The object of this proposal is to assess the relative treatment efficacy of brain surgery versus antiepileptic drug therapy for the treatment of intractable epilepsy. Efficacy will be judged by two criteria: (1) The reduction of seizures during the study for each patient, and (2) The change in functional ability, or "quality of life" as determined by neuropsychological and psychosocial assessments. It is important to assess intellectual, vocational, and psychological functioning (i.e. quality of life) for these patients independently of their seizure control as it cannot be assumed that there will be an association between a reduction of seizures and improvements in objective and subjective indicators of quality of life. No prospective randomized clinical trial has previously been done to answer either of these questions. The methods of the study will be to prospectively identify 100 patients who have medically intractable complex partial epilepsy, and randomly assign them to either medical or surgical treatment. The medical arm will provide the best non-experimental treatment available at a university epilepsy center. Patients will receive antiepileptic drug therapy for 18 months to assess seizure control and quality of life. The surgical arm also will provide the best non-experimental treatment. Patients will undergo a thorough presurgical evaluation, including the use of depth electrodes, and then be offered epilepsy surgery. Surgical patients will be followed and maintained on antiepileptic drugs for 18 months to assess seizure control and quality of life. The entire study will take five years. The practice of epilepsy surgery is expanding beyond the confines of academic hospitals. The priorities of these hospitals on profit may affect the patient selection criterion, the type and depth of workup, and the outcomes. Before this evolution occurs, and the expenditure of hundreds of millions of health care dollars, it is necessary to ask whether the proposed treatment (epilepsy surgery) is both efficacious and cost effective. Thus, this study will calculate the costs of both treatments, as well as the costs and resources used by the patients outside the treatment arms for the full 18 months. A cost/benefit analysis will then be performed.